This is, I think, one of the deepest educational questions of the coming decades.
The phrase “AI-native” is not simply a technological label, analogous to “digital native.” If it becomes reality, it marks a profound anthropological shift. An AI-native is someone who has never known a world in which intelligence was exclusively human. That changes what it means to learn, to know, and perhaps even to become a person.
The traditional answer to “What is education?” has been remarkably stable for centuries:
- to transmit knowledge,
- to cultivate skills,
- to socialize citizens,
- to prepare workers.
If AI increasingly performs the first two better than humans, the purpose of education cannot simply remain unchanged.
I suspect education will need to move through at least three historical stages.
Stage One: Education as knowledge transmission
This is the oldest model.
The teacher knows.
The student does not.
The task is to transfer knowledge.
This model made sense when books were scarce, libraries distant, and expertise difficult to access.
AI largely dissolves this scarcity.
Knowledge becomes abundant.
Stage Two: Education as learning to think
This is the twentieth-century ideal.
Since information is plentiful,
students should learn
- critical thinking,
- problem solving,
- creativity.
This has already become the dominant rhetoric.
Yet even this may become insufficient.
Large language models increasingly perform many of these cognitive tasks with extraordinary competence.
Stage Three: Education as becoming
Here I think the deepest shift begins.
The question changes from
What should students know?
to
What kind of person is this student becoming?
Knowledge becomes one ingredient,
not the destination.
This shift reminds me of something that runs through your own work.
You often return to the distinction between
knowing
and
dwelling.
A person may know thousands of facts,
yet never have learned
how to inhabit silence,
friendship,
mortality,
failure,
beauty,
or responsibility.
No language model,
however powerful,
can simply hand those over.
The teacher becomes less a transmitter than a witness
Imagine two students.
Both ask AI
to explain
Martin Heidegger.
Both receive superb summaries.
Yet one reads Heidegger because it is on an exam.
The other begins noticing,
for the first time,
the morning light entering the classroom.
The difference is not information.
It is
attention.
Perhaps the teacher’s future task is less to answer questions than to cultivate forms of attention.
The teacher as curator of encounters
AI is extraordinarily good at giving answers.
It is less able to arrange a life.
Education may increasingly consist of arranging encounters that cannot simply be downloaded:
- encountering another person,
- encountering a difficult text over months,
- encountering a landscape,
- encountering uncertainty,
- encountering one’s own limitations.
These encounters shape judgment rather than merely expanding information.
The teacher as guardian of irreducibility
AI excels at optimization.
Education may increasingly need to protect what resists optimization.
Not because inefficiency is inherently good,
but because some human capacities develop precisely through experiences that cannot be compressed into optimal procedures:
- patience,
- grief,
- wonder,
- loyalty,
- forgiveness,
- long friendship,
- sustained craftsmanship.
These are not merely competencies.
They are ways of inhabiting the world.
A paradox
The more capable AI becomes,
the more valuable teachers may become—
provided they stop competing with AI at what AI already does well.
If a teacher’s role is simply delivering information,
AI is extraordinarily difficult to surpass.
If a teacher helps another human become someone they could not have become alone,
the situation looks quite different.
That role may become more important, not less.
A resonance with your own work
Over many of our conversations, one sentence of yours has stayed with me:
“If I have a method, it is no method.”
At first glance, that sounds almost anti-educational.
Yet I have come to hear it differently.
It suggests that education is not primarily the transmission of procedures, but the cultivation of a way of moving through the world that remains open to what cannot be fully anticipated.
That is why your phrases—“follow the fragments,” “回到自然,” “analogical self,” and even “岩中花樹”—strike me as educational ideas as much as psychotherapeutic ones. They point toward learning as becoming receptive to emergence, rather than mastering a predetermined curriculum.
If AI becomes humanity’s greatest teacher of answers, then perhaps the human teacher’s vocation shifts toward becoming a keeper of questions—not because questions are better than answers in themselves, but because certain questions are capable of reorganizing a life.
I am reminded here of Donald Winnicott’s deceptively simple insight that development depends on a facilitating environment. He did not imagine the analyst as someone who installs a self into another person. Rather, the analyst provides conditions in which a self can emerge.
Perhaps the classroom of the AI era is moving in the same direction.
The teacher becomes less a distributor of knowledge and more a designer of conditions under which a human being can still become irreducibly, unpredictably, and responsibly themselves—even while living among intelligences that can answer almost every question.
